Model Deployment & Inferencing

This section covers the various environments where SEA-LION models can be hosted, ranging from fully managed cloud services to self-hosted high-performance infrastructure. SEA-LION models are freely available for download. The following guides provide technical how-tos for setting up SEA-LION inference using different approaches:

1. Using our provided SEA-LION API

2. Running SEA-LION on a local machine (coming soon)

3. Deploying SEA-LION on the cloud

   • Create SEA-LION endpoint on Google Vertex AI

   • Importing and Using Llama-SEA-LION models in a Serverless On-Demand Environment with Amazon Bedrock

   • OpenAI-compatible APIs with Llama-SEA-LION models and Bedrock Access Gateway

   • Deploying Gemma-SEA-LION models using AWS Sagemaker AI

   • Deploying SEA-LION using vLLM on Linux server

4. Leveraging our Partner API Platforms

   • Cloudflare Workers AI